Methodology for Bayesian monotonic polynomials

Andrew Manderson

    Research output: ThesisMaster's Thesis

    522 Downloads (Pure)

    Abstract

    Many physical processes are postulated to obey a monotonic relationship, whereby the entity of interest must strictly increase or decrease as a function of a covariate. Monotonic polynomials are a popular tool for incorporating such a priori knowledge into statistical models, particularly in settings where irreducible noise induces unconstrained model fits which violate the presupposed monotonic relationship. This thesis develops novel Bayesian methodologies for fitting monotonic polynomials and selecting the appropriate polynomial model, in a variety of scenarios.
    Original languageEnglish
    QualificationMasters
    Awarding Institution
    • The University of Western Australia
    Supervisors/Advisors
    • Turlach, Berwin, Supervisor
    • Murray, Kevin, Supervisor
    Thesis sponsors
    Award date13 Jun 2018
    DOIs
    Publication statusUnpublished - 2018

    Fingerprint

    Dive into the research topics of 'Methodology for Bayesian monotonic polynomials'. Together they form a unique fingerprint.

    Cite this